Intelligent video monitoring system based on holographic image

Abstract

The invention relates to the technical field of video monitoring, in particular to an intelligent video monitoring system based on holographic images, which comprises a multispectral imaging module, an abnormal energy correction module, a multispectral dynamic monitoring module and a behavior track analysis module. In the invention, through capturing light wave amplitude, frequency and phase information and deeply analyzing light wave energy and spatial distribution, the perception capability of spectral characteristics and energy change in a monitoring area is improved, abnormal high-value energy distribution is identified and parameters are adjusted to realize accurate correction, and detection and correction of abnormal light wave distribution are enhanced. Based on light wave frequency change and dynamic target characteristic matching, the recognition capability of a dynamic target in a complex scene is improved, a target trajectory is comprehensively tracked, the behavior dynamic state is accurately judged, multispectral information analysis and multidimensional dynamic analysis are combined, the precise recognition, abnormal analysis and dynamic behavior interpretation capabilities of monitoring in a complex environment are remarkably enhanced, and the monitoring accuracy is improved. And efficient support is provided for intelligent video monitoring.

Description

Technical Field

[0001] The present invention relates to the field of video surveillance technology, and in particular to an intelligent video surveillance system based on holographic imaging. Background Art

[0002] The field of video surveillance technology includes technologies for monitoring and recording objects, behaviors, events, etc. in a scene. Its core content is to use video acquisition equipment to obtain real-time images of the target area, and to monitor and manage the target by processing, analyzing and storing video data. This technical field covers image acquisition, video encoding, transmission, storage, target detection and analysis, and is widely used in public safety, traffic management, industrial monitoring and other fields. With the development of technology, video surveillance has gradually evolved from traditional analog monitoring to digital and intelligent directions. By introducing artificial intelligence and big data analysis technologies, the automation and intelligence level of video surveillance can be further improved.

[0003] Among them, the intelligent video surveillance system refers to a monitoring system that performs real-time analysis and judgment of video data based on computer vision and video processing technology. The technical matters targeted by the patent subject include obtaining video data through image acquisition equipment in the monitoring system, using video processing technology to parse images and scenes in the data, and extracting key information based on pattern recognition and target detection technology. The system establishes an image feature database for specific scenes and combines image matching technology to achieve accurate identification and behavior analysis of target objects. The overall system is based on video data and completes the monitoring function through image parsing and analysis technology.

[0004] The existing technology relies on image features within the visible spectrum to identify abnormal targets in video surveillance, but lacks the ability to fully perceive a wider spectrum and light wave characteristics (such as frequency, amplitude, etc.), resulting in insufficient recognition accuracy for non-explicit features. In dynamic target monitoring, traditional methods are mostly achieved through the simple superposition of continuous frame image differences or motion vectors, making it difficult to accurately analyze subtle trajectory changes and complex motion patterns of targets, especially in scenes with drastic changes in illumination or with many target occlusions, which can easily lead to misjudgments or omissions. In the detection and correction of abnormal energy distribution, existing systems are mostly based on a single threshold judgment method, which has poor adaptability to complex energy abnormality scenes and cannot efficiently respond to energy fluctuations in a changing environment. In terms of behavior analysis, there is a lack of dynamic adaptation and prediction capabilities for unknown behavior trajectories, resulting in insufficient comprehensiveness and accuracy for target behaviors, which limits the monitoring capabilities of existing technologies in complex environments, dynamic scenes, and multi-spectral information processing, affecting the effectiveness of monitoring in high-precision, multi-dimensional applications. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent video monitoring system based on holographic imaging.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: An intelligent video monitoring system based on holographic imaging comprises:

[0007] The multispectral imaging module is based on a multispectral holographic imaging device, which captures light wave amplitude, light wave frequency and light wave phase information, extracts infrared, ultraviolet and visible spectrum data, analyzes light wave energy and spatial distribution information, and generates a light wave energy distribution data set;

[0008] The abnormal energy correction module analyzes the high value of light wave energy and the fluctuation amplitude of the partition based on the light wave energy distribution data set, matches the abnormal high value distribution area, identifies the light wave abnormal distribution area, obtains the abnormal energy identification data set, adjusts the distribution parameters in combination with the light wave characteristic differences in the monitoring image, and generates the corrected energy distribution result;

[0009] The multi-spectral dynamic monitoring module extracts the light wave frequency characteristics of the frequency change area based on the corrected energy distribution result, analyzes the light wave frequency change area in combination with the dynamic target characteristics in the monitoring picture frame, generates the frequency change distribution result, matches the light wave characteristics and the video monitoring target characteristics in the area, locates the multi-spectral dynamic feature area, and obtains the multi-spectral dynamic area information;

[0010] The behavior trajectory analysis module analyzes the spatial displacement and movement direction of the light wave amplitude in continuous frames based on the multi-spectral dynamic area information, generates a set of motion vectors of the target area in combination with the spatial change trajectory, performs time series analysis on the motion vector points, extracts trajectory change nodes, and generates dynamic behavior recognition results of the monitored target trajectory through node connection and trajectory fitting.

[0011] As a further solution of the present invention, the steps of acquiring the light wave energy distribution data set are specifically:

[0012] Based on multispectral holographic imaging equipment, the light wave amplitude, frequency and phase information are captured, infrared, ultraviolet and visible spectrum data are extracted, and the basic network structure of light wave energy distribution is established;

[0013] The spatial distribution information of the basic network structure of light wave energy distribution is used, combined with the band weight parameters corresponding to the infrared, ultraviolet and visible spectrum data, to calculate the regional light wave energy value using the formula:

[0014]

[0015] Generate regional light wave energy analysis table;

[0016] Among them, E represents the regional light wave energy value, Ai represents the amplitude energy of the i-th light wave, P i represents the phase energy of the i-th light wave, W i represents the band weight parameter corresponding to the i-th light wave, x i Represents the horizontal axis of spatial distribution, y i represents the ordinate of spatial distribution, K is the adjustment coefficient, and n represents the total number of light waves;

[0017] Based on the regional light wave energy analysis table, the light wave energy values ​​of the differentiated areas are compared, the energy distribution characteristics are analyzed, the spatial distribution trend of the light wave energy and the energy difference information are identified, and the light wave energy distribution data set is generated.

[0018] As a further solution of the present invention, the step of acquiring the abnormal energy recognition data set is specifically:

[0019] Calling the light wave energy distribution data set, dividing the light wave energy value into intervals by analyzing the light wave energy value and the partition fluctuation amplitude in the area, marking the spatial distribution coordinates of the light wave energy value area, and generating a light wave energy high value spatial distribution set;

[0020] Based on the high-value spatial distribution set of light wave energy, the energy difference in the high-value area is analyzed, and combined with the partition fluctuation amplitude parameter, the formula is adopted:

[0021]

[0022] Match the abnormal high energy value area to generate abnormal high value matching results;

[0023] Among them, D represents an abnormally high matching value, E j represents the light wave energy value in the jth region, μ represents the mean value of the energy value, V j represents the fluctuation range of the jth region, σ represents the standard deviation of the fluctuation range, β is the smoothing coefficient, and m represents the number of regions;

[0024] The abnormal high value matching result is called, and the abnormal matching value of the area is compared, and combined with the spatial distribution coordinates, the abnormal distribution area of ​​the light wave is identified, the light wave energy and spatial information in the abnormal area are extracted, and the abnormal energy identification data set is generated.

[0025] As a further solution of the present invention, the step of obtaining the corrected energy distribution result is specifically as follows:

[0026] Based on the abnormal energy identification data set, the original data of energy values ​​are extracted, grouped and processed according to the differentiated energy intervals in the data, the frequency of energy values ​​in each interval is counted, and the energy distribution characteristic parameters are obtained by quantifying the cumulative total amount of frequency data and combining the distribution deviation ratio of the energy values ​​in the interval;

[0027] Based on the energy distribution characteristic parameters, light wave characteristic data of the monitoring image is collected, light wave intensity values ​​are allocated to corresponding distribution positions according to energy intervals, the proportional weights of intensity and interval parameters in the light wave data are adjusted, the frequency range of the energy characteristic parameters is allocated, and the light wave distribution parameters after adjustment are obtained;

[0028] Based on the adjusted distribution parameters of the light waves, the energy values ​​are redistributed according to the interval frequencies, the weight ratios of the energy intervals are integrated, the energy value distribution is corrected according to the cumulative contribution of the light waves in the intervals, and a corrected energy distribution result is constructed.

[0029] As a further solution of the present invention, the step of obtaining the frequency change distribution result is specifically:

[0030] Based on the corrected energy distribution result, the energy distribution is decomposed into a frequency change sequence according to the time series, the frequency change rate in the time series is analyzed, and the frequency change area of ​​the light wave is identified in combination with the statistical parameters of the frequency amplitude change range, and the frequency change area marking result is generated;

[0031] The frequency change region marking result is called, the light wave frequency characteristics in the frequency change region are analyzed, and the frequency change characteristic value in the frequency change region is calculated using the formula:

[0032]

[0033] Generate frequency variation characteristic analysis results;

[0034] Among them, F c Represents the frequency change characteristic value within the frequency change area, f k is the light wave frequency value at the kth time point, Δa k is the frequency amplitude change at the kth time point, Q k is the weight parameter, Δf k is the frequency change at the kth time point, α is the smoothing parameter;

[0035] The frequency change characteristic analysis result is called, the frequency change region characteristic and the dynamic target feature are correlated and analyzed, the spatial position of the dynamic target and the intersection of the frequency change region are matched, and the frequency change distribution result is generated.

[0036] As a further solution of the present invention, the step of acquiring the multi-spectral dynamic area information is specifically as follows:

[0037] Based on the frequency change distribution results, the data are grouped according to the spatial region position, the amplitude and change trend of each group of frequency fluctuations are analyzed, the frequency dynamic distribution characteristics of the time axis data are identified, and the distribution correlation parameters between the frequency and the region are integrated to obtain the frequency and region matching data;

[0038] Based on the frequency and area matching data, extract the intensity change characteristics of the differentiated areas in the light wave characteristic data, analyze the dynamic range of the light wave intensity on the time axis, compare the distribution of each set of intensity data with the spatial area, adjust the proportional relationship of the intensity distribution in the light wave characteristic, and generate a light wave and target area matching parameter set;

[0039] Based on the light wave and target area matching parameter set, the spatial position of the multispectral feature is located in the monitoring data, the distribution data of the light wave characteristics in the time axis and the spatial area are analyzed, the multispectral feature frequency distribution weight of the feature area is screened, and the multispectral dynamic area information is obtained.

[0040] As a further solution of the present invention, the step of acquiring the motion vector set of the target area is specifically:

[0041] Based on the multi-spectral dynamic area information, the time series of the light wave amplitude of the continuous frames is called, the spatial displacement of the light wave amplitude is analyzed, and the spatial change amount and directional change amount of the regional light wave amplitude between the continuous frames are analyzed to obtain the displacement distribution characteristics of the light wave amplitude;

[0042] The displacement distribution characteristics of the light wave amplitude are called, combined with the spatial change trajectory of continuous frames, and the movement direction and displacement characteristics of the light wave amplitude are analyzed through trajectory mapping of dynamic targets, using the formula:

[0043]

[0044] Calculate the characteristic value of the light wave motion amplitude to obtain the light wave motion amplitude characteristic;

[0045] Among them, M represents the characteristic value of the amplitude of light wave motion, (x a ,y a ) and (x b ,y b ) represent the starting coordinates and ending coordinates of the target area in the continuous frames, θ1 and θ2 are the angles of the light wave motion direction in the continuous frames, and γ is the weight parameter;

[0046] The light wave motion trajectory feature is called, the spatial displacement of the light wave is combined with the motion direction, and matched with the trajectory parameters of the dynamic target. By analyzing the difference in the trajectory in the matching area, the dynamic motion feature of the target area is extracted, and a motion vector set of the target area is generated.

[0047] As a further solution of the present invention, the steps for obtaining the dynamic behavior recognition result of the monitoring target trajectory are specifically as follows:

[0048] Based on the motion vector set of the target area, the vector points are arranged in time order, the spatial displacement and time interval of adjacent vector points are identified, the displacement and time change ratio are compared, the dynamically changing vector points are screened, and a trajectory change node set is generated;

[0049] Based on the trajectory change node set, the spatial distance and direction angle of the displacement path between adjacent nodes are identified, the path is converted into a curve for trajectory fitting, the continuity and local change rate of the fitting curve are parameterized and analyzed, and a trajectory fitting result is generated;

[0050] Based on the trajectory fitting results, the speed and direction changes in the time series are analyzed in segments, the key positions of the direction changes are marked, the action nodes are extracted, and the dynamic behavior recognition results of the monitoring target trajectory are generated.

[0051] Compared with the prior art, the advantages and positive effects of the present invention are:

[0052] In the present invention, by capturing light wave amplitude, light wave frequency and light wave phase information, and conducting in-depth analysis of light wave energy and spatial distribution information, the perception ability of spectral characteristics and energy changes in the monitoring area is significantly improved, the abnormal area of ​​light wave energy distribution is analyzed, the fluctuation amplitude and distribution law of abnormal high-value energy are identified, and more accurate energy anomaly correction is achieved by adjusting energy distribution parameters, thereby enhancing the detection and correction ability of abnormal light wave distribution in the environment. Based on light wave frequency changes and dynamic target characteristic matching, the dynamic feature area in the monitoring screen can be more efficiently identified, multi-spectral dynamic changes can be located, and the accurate recognition ability of dynamic targets in complex scenes can be improved. Combined with the spatial displacement and motion direction analysis of light wave amplitude in continuous frames, the comprehensive tracking of the motion trajectory of the monitoring target and the accurate judgment of the behavior dynamics can be achieved through trajectory node extraction and fitting, and the monitoring ability of understanding and predicting the target behavior can be improved. In the processing logic, through the deep analysis of multi-spectral information and multi-dimensional dynamic analysis, the organic combination of target characteristics, light wave energy characteristics and dynamic trajectory is achieved, and the accurate recognition, abnormal analysis and dynamic behavior interpretation capabilities of monitoring in complex environments are significantly improved, providing more efficient technical support for the intelligentization of video monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a system flow chart of the present invention;

[0054] Figure 2 It is a flow chart of the light wave energy distribution data set in the present invention;

[0055] Figure 3 A flowchart of an abnormal energy identification data set in the present invention;

[0056] Figure 4It is a flow chart of the energy distribution result after correction in the present invention;

[0057] Figure 5 It is a flow chart of the frequency change distribution result in the present invention;

[0058] Figure 6 It is a flow chart of multi-spectral dynamic area information in the present invention;

[0059] Figure 7 A flow chart of a motion vector set of a target area in the present invention;

[0060] Figure 8 It is a flow chart of the dynamic behavior recognition results of monitoring target trajectory in the present invention. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0062] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0063] See also Figure 1 , an intelligent video surveillance system based on holographic imaging includes:

[0064] The multispectral imaging module is based on a multispectral holographic imaging device, which captures light wave amplitude, light wave frequency and light wave phase information, extracts infrared, ultraviolet and visible spectrum data, analyzes light wave energy and spatial distribution information, and generates a light wave energy distribution data set;

[0065] The abnormal energy correction module is based on the light wave energy distribution data set, analyzes the light wave energy high value and the partition fluctuation amplitude, matches the abnormal high value distribution area, identifies the light wave abnormal distribution area, obtains the abnormal energy identification data set, adjusts the distribution parameters based on the light wave characteristic differences in the monitoring image, and generates the corrected energy distribution result;

[0066] Based on the corrected energy distribution results, the multi-spectral dynamic monitoring module extracts the light wave frequency characteristics of the frequency change area, combines the dynamic target characteristics in the monitoring picture frame, analyzes the light wave frequency change area, generates the frequency change distribution results, matches the light wave characteristics and video monitoring target characteristics in the area, locates the multi-spectral dynamic feature area, and obtains the multi-spectral dynamic area information;

[0067] The behavior trajectory analysis module analyzes the spatial displacement and movement direction of the light wave amplitude in continuous frames based on multi-spectral dynamic area information, combines the spatial change trajectory, generates a set of motion vectors for the target area, performs time series analysis on the motion vector points, extracts trajectory change nodes, and generates dynamic behavior recognition results of the monitored target trajectory through node connection and trajectory fitting.

[0068] The light wave energy distribution data set includes light wave energy intensity distribution, light wave spatial distribution and light wave energy partition data. The abnormal energy identification data set includes abnormal high value area information, partition fluctuation amplitude analysis results and abnormal area matching results. The corrected energy distribution results include the corrected energy distribution diagram, adjusted distribution parameters and partition correction energy values. The frequency change distribution results include light wave frequency change area distribution, light wave frequency change amplitude and dynamic frequency characteristic data. The multi-spectral dynamic area information includes the dynamic target area position, multi-spectral characteristic area range and light wave dynamic characteristic information. The motion vector set of the target area includes the direction of the motion vector, the size of the motion vector and the time series distribution data. The dynamic behavior recognition results of the monitoring target trajectory include trajectory change nodes, trajectory fitting results and target dynamic behavior characteristics.

[0069] See also Figure 2 , the steps for obtaining the light wave energy distribution data set are as follows:

[0070] Based on multispectral holographic imaging equipment, the light wave amplitude, frequency and phase information are captured, infrared, ultraviolet and visible spectrum data are extracted, and the basic network structure of light wave energy distribution is established;

[0071] When calling the multispectral holographic imaging device to capture band parameters, the original light wave signal collected by the sensor needs to be decomposed into infrared, ultraviolet and visible spectral bands. The frequency and phase parameters of the light wave are calculated, and the infrared spectrum data is subjected to Fourier transform to extract its frequency distribution. The energy amplitude is obtained by calculating the energy density function of the ultraviolet spectrum data. For the visible spectrum data, its amplitude and phase parameters are called for partition sampling through the spectrometer, and then the light wave amplitude energy and phase energy of each partition are weightedly calculated respectively. The weighted value is based on the relative energy contribution of each band in the actual monitoring process. The comprehensive distribution of the light wave amplitude energy and phase energy is calculated, and finally the basic network structure of light wave energy distribution is established.

[0072] The spatial distribution information of the basic network structure of light wave energy distribution is used, combined with the band weight parameters corresponding to the infrared, ultraviolet and visible spectrum data, to calculate the regional light wave energy value using the formula:

[0073]

[0074] Generate regional light wave energy analysis table;

[0075] Among them, E represents the regional light wave energy value, A i represents the amplitude energy of the i-th light wave, P i represents the phase energy of the i-th light wave, W i represents the band weight parameter corresponding to the i-th light wave, x i Represents the horizontal axis of spatial distribution, y i represents the ordinate of spatial distribution, K is the adjustment coefficient, and n represents the total number of light waves;

[0076] The benefit of the formula is that by introducing the band weight parameter W i As well as the correction of energy due to spatial coordinate differences, the comprehensive effects of light wave energy on amplitude, phase and spatial distribution are comprehensively considered, so that the distribution characteristics of the comprehensive energy of light waves can be described more accurately;

[0077] E represents the comprehensive energy of regional light waves, A i represents the energy of the i-th light wave amplitude, which is obtained by calculating the light wave amplitude energy density function, P i represents the phase energy of the i-th light wave, obtained by Fourier transform, W i represents the band weight parameter corresponding to the i-th light wave, which is calculated by the relative contribution of each band energy to the overall light wave energy. i Represents the horizontal axis of spatial distribution, y i The ordinate represents the spatial distribution, which is obtained through the spatial calibration parameters of the sensor in the light wave capture device. K is the adjustment coefficient, which is obtained based on the balance adjustment of the actual energy distribution.

[0078] Values: A1=5, A2=4.5, P1=0.8, P2=0.7, W1=1.2, W2=1.1, x1

[0079] =2, y1=1.8, x2=3, y2=2.2, K=0.5;

[0080] Calculate E:

[0081]

[0082] E = 5.07 + 2.48 = 7.55;

[0083] The result shows that the calculated light wave energy value is 7.55, which reflects the total energy value under the current light wave amplitude, phase and spatial weight distribution conditions, and can be further used as a basis for distribution feature extraction and regional light wave energy analysis.

[0084] Based on the regional light wave energy analysis table, compare the light wave energy values ​​of differentiated areas, analyze the energy distribution characteristics, identify the spatial distribution trend and energy difference information of light wave energy, and generate a light wave energy distribution data set;

[0085] By comparing the comprehensive energy values ​​of light waves in different regions and extracting the comprehensive energy differences of light waves, it is necessary to normalize the comprehensive energy values ​​of each region, and calculate the maximum, minimum and standard deviation values ​​through the statistical distribution of the normalized values ​​to determine whether the distribution characteristics of the comprehensive energy of light waves are uniform. By defining the difference threshold, the unevenly distributed areas are determined, and the normalized comprehensive energy values ​​are compared one by one with the theoretical energy values ​​under the uniform distribution condition. For areas where the difference value exceeds the defined threshold, they are marked as unevenly distributed areas. At the same time, the comprehensive energy matrix is ​​called to extract the energy parameters of the area for further distribution feature analysis. Finally, the spatial distribution trend and energy difference information of light wave energy are extracted to generate a light wave energy distribution data set.

[0086] See also Figure 3 ,The specific steps for obtaining the abnormal energy recognition data set are:

[0087] Call the light wave energy distribution data set, divide the light wave energy value into intervals by analyzing the light wave energy value and the partition fluctuation amplitude in the area, mark the spatial distribution coordinates of the light wave energy value area, and generate a light wave energy high value spatial distribution set;

[0088] First, the data set needs to be divided into several regional distribution units, and the joint distribution parameters of amplitude and frequency are used to extract high energy values. The difference in light wave energy distribution is quantified as the ratio of amplitude to fluctuation standard deviation by statistically analyzing the standard deviation of the partition fluctuation amplitude. Then, each regional distribution unit is screened. After the screening is completed, the high-value data of the regional distribution unit is called and the quantitative ratio of its fluctuation difference is calculated. The area where the ratio exceeds the specified threshold is marked as a high-value light wave energy area. By combining the spatial identification parameters of the area and the statistical mean of the distribution unit, a high-value spatial distribution set of light wave energy is generated.

[0089] Based on the high-value spatial distribution set of light wave energy, the energy difference in the high-value area is analyzed, and combined with the partition fluctuation amplitude parameter, the formula is adopted:

[0090]

[0091] Match the abnormal high energy value area to generate abnormal high value matching results;

[0092] Among them, D represents an abnormally high matching value, E j represents the light wave energy value in the jth region, μ represents the mean value of the energy value, V j represents the fluctuation range of the jth region, σ represents the standard deviation of the fluctuation range, β is the smoothing coefficient, and m represents the number of regions;

[0093] The benefit of the formula is that by combining the absolute value of the energy difference, the standard deviation of the fluctuation amplitude, and the smoothing adjustment coefficient, it can effectively screen out energy anomaly areas, while weighing the spatial volatility and local deviation characteristics of energy distribution, and improving the accuracy and stability of abnormal energy identification;

[0094] D represents the abnormally high value matching result, which is obtained by summing up, E j represents the light wave energy value of the jth region, which is directly called from the light wave energy distribution dataset. μ represents the mean value of the light wave energy, which is calculated by the statistical mean of the energy values ​​of all regions in the dataset. V j represents the fluctuation amplitude of the jth region, which is calculated by the amplitude standard deviation of the light wave energy time series in the region. σ represents the overall standard deviation of the fluctuation amplitude, which is obtained by the statistical analysis of the fluctuation amplitude of the light wave energy distribution data set. β is the smoothing coefficient, and its specific value is adjusted according to the balance of light wave energy distribution to avoid the interference of small fluctuations on abnormal energy matching results.

[0095] Parameter value:

[0096] E1=150, E2=175, μ=160, V1=10, V2=15, σ=8, β=2;

[0097] Calculate D:

[0098]

[0099] D = 10 + 22.5 = 32.5;

[0100] The result shows that the calculated abnormal high value match value is 32.5, indicating that there is a significant deviation between the light wave energy in the area and the overall mean. Combined with the spatial distribution parameters, the abnormal area can be further located for subsequent processing.

[0101] The abnormal high value matching results are called, and the abnormal matching values ​​of the regions are compared and combined with the spatial distribution coordinates to identify the abnormal distribution areas of light waves, extract the light wave energy and spatial information in the abnormal areas, and generate abnormal energy identification data sets;

[0102] When comparing the abnormal matching values ​​of each area, it is necessary to combine the distribution of the abnormal matching values ​​with the regional spatial coordinate information, divide the matching values ​​through the statistical distribution threshold, and partition the abnormal value distribution and the normal distribution through the difference ratio formula. Then, the abnormal high value parameters of the area are called to map the spatial coordinates. The overall distribution characteristics of the optical wave abnormal area are analyzed through the spatial mapping matrix of the abnormal distribution, and the distribution characteristic parameters are matched and integrated with the spatial coordinates to extract the optical wave energy data and position parameters of the abnormal area, and finally generate an abnormal energy recognition data set.

[0103] See also Figure 4 , the specific steps for obtaining the corrected energy distribution result are:

[0104] Based on the abnormal energy recognition data set, the original data of energy values ​​are extracted, grouped according to the differentiated energy intervals in the data, and the energy value frequency of each interval is counted. The energy distribution characteristic parameters are obtained by quantifying the cumulative total amount of frequency data and combining the distribution deviation ratio of the energy value in the interval.

[0105] According to the distribution of energy values ​​in different intervals, multiple intervals are divided for grouping processing. In the specific implementation process, the frequency value of each energy interval is first counted, and the cumulative total of each interval is obtained by cumulative calculation. The proportion of energy values ​​in each interval is quantified by calculating the ratio of the cumulative total to the total amount of overall data interval by interval. The energy value in the interval is offset calculated based on the distribution characteristics of the energy value. Based on the frequency distribution of the energy value, the center of gravity offset rate of the energy value in each interval is gradually calculated. The cumulative offset ratio is set to measure the distribution balance in the interval. The energy distribution characteristic parameter set is generated by superimposing the frequency proportion and the offset ratio. The distribution differences of each energy interval in the data are described based on this parameter set.

[0106] Based on the energy distribution characteristic parameters, the light wave characteristic data of the monitoring image is collected, the light wave intensity value is allocated to the corresponding distribution position according to the energy interval, the proportional weight of the intensity and interval parameters in the light wave data is adjusted, the frequency range of the energy characteristic parameters is allocated, and the light wave distribution parameters after adjustment are obtained;

[0107] Combined with the characteristic parameters corresponding to each energy interval, the intensity value of the light wave characteristic data is allocated to the corresponding energy interval distribution position. In actual operation, the light wave intensity in the monitoring image data is quantized and decomposed into continuous light wave intensity intervals. The energy distribution parameters are matched interval by interval. The proportional relationship between the light wave intensity value and the energy interval is established through the normalization operation of the intensity value. The proportional weight of the intensity and interval parameters in the light wave intensity data is adjusted. The light wave characteristic data is redistributed based on the frequency range of the energy interval. By adjusting the weight ratio, the light wave energy frequency distribution value of each interval is generated. After integrating the data, the adjusted light wave distribution parameters are obtained, thereby providing a basis for subsequent energy redistribution.

[0108] Based on the adjusted distribution parameters of the light waves, the energy values ​​are redistributed according to the interval frequencies, the weight ratios of the energy intervals are integrated, the energy value distribution is corrected according to the cumulative contribution of the light waves in the intervals, and the corrected energy distribution results are constructed;

[0109] The energy values ​​are redistributed according to the interval frequencies, and the weight proportions within the energy intervals are integrated in turn. The adjusted distribution parameters of the light waves are taken as input, and the cumulative contribution rates of the interval energy are calculated one by one. The weights of the energy intervals are corrected by superimposing the contribution rate values ​​of the light wave data within the intervals. The energy values ​​are redistributed according to the corrected weights, and a cumulative distribution function of the energy values ​​is constructed. The energy distribution proportion within the intervals is gradually adjusted, and the cumulative value of the light wave contribution rate is used to construct the final correction result, which reflects the correction effect of the cumulative contribution of light waves in each energy interval on the overall energy distribution, and generates a corrected energy distribution result to optimize the energy recognition capability of the intelligent video surveillance system for holographic images.

[0110] See also Figure 5 , the specific steps for obtaining the frequency change distribution results are:

[0111] Based on the corrected energy distribution results, the energy distribution is decomposed into a frequency change sequence according to the time series, the frequency change rate in the time series is analyzed, and the frequency change area of ​​the light wave is identified by combining the statistical parameters of the frequency amplitude change range, and the frequency change area marking result is generated;

[0112] When extracting the light wave frequency characteristics of the frequency change area, it is necessary to decompose the corrected energy distribution results into time series, detect the frequency change trend through the amplitude change of the light wave time series, calculate the mean, variance and standard deviation of the frequency change in each area, and screen out the area with higher variance and mark it as the frequency change area. After marking, the amplitude of the frequency series in the area is counted through the identification of the frequency change area, and the maximum and minimum values ​​of its amplitude change are called to calculate the frequency change interval. At the same time, combined with the differential characteristics of the time series, the frequency change rate is extracted as a dynamic characteristic parameter, and finally the frequency change area marking result is generated.

[0113] Call the frequency change area marking result, analyze the light wave frequency characteristics in the frequency change area, and calculate the frequency change characteristic value in the frequency change area using the formula:

[0114]

[0115] Generate frequency variation characteristic analysis results;

[0116] Among them, F c Represents the frequency change characteristic value within the frequency change area, f k is the light wave frequency value at the kth time point, Δa k is the frequency amplitude change at the kth time point, Q k is the weight parameter, Δf k is the frequency change at the kth time point, α is the smoothing parameter;

[0117] The benefit of the formula is that by introducing the frequency amplitude change at the time point, the weight parameter and the square sum characteristic of the frequency change, the dynamic change of the frequency change area can be weighted and corrected, thereby improving the accuracy and sensitivity of the frequency change analysis results;

[0118] F c represents the comprehensive characteristics of frequency change, which is obtained by calculating the dynamic change characteristics of the time point after weight correction, f k is the light wave frequency value at the kth time point, which is directly obtained through the corrected energy distribution time series, Δa k is the frequency amplitude change at the kth time point, obtained by calculating the difference between adjacent points in the amplitude sequence, Q k is the weight parameter of the kth time point, which is set after evaluating the importance of dynamic changes at different time points to the analysis results, Δf k is the frequency change at the kth time point, obtained by the difference of the frequency series, and α is the smoothing parameter, which is used for correction calculation after comprehensive analysis of the sequence change characteristics;

[0119] Parameter values: f1=10, f2=15, f3=12, Δa1=2, Δa2=1.5, Δa3=2.2, Q1=1.2, Q2=1.0, Q3=1.1, Δf1=5, Δf2=3, Δf3=4, α=0.5;

[0120] Calculate F c :

[0121]

[0122] The results show that the frequency change characteristic value in the frequency change region is 10.63, indicating that the intensity of the change in the light wave frequency in this region is relatively significant, which can serve as an important reference for subsequent dynamic characteristic analysis.

[0123] The frequency change characteristic analysis results are called, the frequency change area characteristics are correlated with the dynamic target characteristics, the spatial position of the dynamic target and the intersection of the frequency change area are matched, and the frequency change distribution results are generated;

[0124] Firstly, the spatial position of the dynamic target feature is calibrated, and the intersection operation is performed on the calibrated position parameters and the spatial distribution information of the frequency change area to determine the associated area between the spatial feature range of the dynamic target and the frequency change area. The motion direction of the dynamic target is analyzed through the light wave frequency change characteristics of the associated area. The change trend of the motion trajectory is calculated by combining the motion direction, the amplitude of the frequency change and the dynamic target position deviation. The area with the most significant frequency change in the trajectory change is extracted, and finally the frequency change distribution result is generated.

[0125] See also Figure 6 ,The specific steps for obtaining multi-spectral dynamic area information are:

[0126] Based on the frequency change distribution results, the data are grouped according to the spatial region location, the amplitude and change trend of each group of frequency fluctuations are analyzed, the frequency dynamic distribution characteristics of the time axis data are identified, and the distribution correlation parameters between frequency and region are integrated to obtain the frequency and region matching data;

[0127] The frequency distribution data in each spatial area is gradually extracted, and its frequency fluctuation amplitude and change trend are counted and calculated. The time axis data is divided into multiple sub-areas according to the spatial area. The frequency change values ​​in each area are counted, and the amplitude of the frequency fluctuation in the area is calculated using the cumulative frequency deviation. The fluctuation trend is quantified according to the amplitude change rate. Combined with the dynamic distribution on the time axis, the distribution characteristics of the frequency values ​​in different time periods are further analyzed. By comparing and analyzing the change ratio of frequency data in different spatial areas, the dynamic frequency distribution characteristics in the area are formed. The distribution correlation parameters of frequency and spatial area are integrated, the frequency changes between areas are weighted, and a matching parameter set of area and frequency distribution is established. Finally, the frequency and area matching data are obtained to support the dynamic analysis of space and frequency characteristics in holographic image monitoring.

[0128] Based on the frequency and area matching data, the intensity change characteristics of the differentiated areas in the light wave characteristic data are extracted, the dynamic range of the light wave intensity on the time axis is analyzed, the distribution of each set of intensity data and the spatial area is compared, the proportional relationship of the intensity distribution in the light wave characteristics is adjusted, and a matching parameter set of the light wave and the target area is generated;

[0129] Combined with the matching results of the frequency distribution in each group of regions, the light wave data is analyzed for intensity in each region. By extracting the light wave intensity variation range on the time axis, the dynamic characteristics of the intensity value in each time period are quantified. The frequency distribution of each group of intensity data and the spatial region are gradually compared, and the distribution offset of the intensity value is analyzed. Combined with the dynamic variation range of the light wave characteristic intensity, the proportion of the intensity distribution in the spatial region is recalculated, and the intensity ratio between different regions is adjusted to form a new intensity distribution ratio parameter. By integrating the parameters, a parameter set for matching the light wave with the target region is generated, which provides a basis for subsequent light wave dynamic monitoring and regional characteristic matching, while improving the accuracy of the holographic imaging monitoring system in target area identification.

[0130] Based on the matching parameter set between the light wave and the target area, locate the spatial position of the multispectral feature in the monitoring data, analyze the distribution data of the light wave characteristics in the time axis and spatial area, screen the multispectral feature frequency distribution weight of the feature area, and obtain the multispectral dynamic area information;

[0131] By matching each set of light wave intensity data with the frequency dynamic characteristics in the area, the specific distribution of multi-spectral features is identified, and the distribution data of light wave characteristics in the time axis and spatial area are gradually analyzed. The distribution density of light waves in each area is calculated by extracting the cumulative value of multi-spectral frequency, and the regional positions with significant characteristics are screened. Combined with the screened feature areas, the distribution weights of the multi-spectral feature frequencies in the area are further calculated, and the weight data are accumulated to obtain the frequency distribution characteristics of the multi-spectral dynamic area. Through this dynamic analysis method, the light wave distribution data of the multi-spectral feature area is located, providing data support for dynamic area analysis for holographic image monitoring, while optimizing the presentation effect of multi-spectral features in the monitoring screen, and obtaining multi-spectral dynamic area information.

[0132] See also Figure 7 , the steps for obtaining the motion vector set of the target area are specifically as follows:

[0133] Based on the multi-spectral dynamic area information, the time series of the light wave amplitude of continuous frames is called to analyze the spatial displacement of the light wave amplitude. The spatial and directional changes of the regional light wave amplitude between continuous frames are analyzed to obtain the displacement distribution characteristics of the light wave amplitude.

[0134] The continuous frames are preprocessed to extract the light wave amplitude distribution of each time frame. The light wave distribution in the frame is spatially segmented by region division and the position parameters of each region are marked. For each segmented region, the light wave amplitude distribution change between continuous frames is extracted according to the time series, the spatial displacement of the light wave amplitude center of each region is calculated, and the displacement characteristics of each region are calculated by using the coordinate change of the amplitude center point, including the amplitude and direction of the displacement. The displacement characteristics of the continuous frames are further statistically analyzed, and the displacement amplitude change characteristics of each region are quantified by calculating the average value and standard deviation of the displacement, and the directional angle of the amplitude center offset is calculated, and the change characteristics of the directional angle are extracted to form the directional offset data of the displacement. Combined with the amplitude trajectory change between frames, the spatial change trajectory of the light wave amplitude in the continuous frames is tracked and processed to form a trajectory distribution. The spatial distribution characteristics are integrated to calculate and merge the trajectories of the regions to obtain the displacement distribution characteristics of the light wave amplitude.

[0135] The displacement distribution characteristics of the light wave amplitude are called, combined with the spatial change trajectory of continuous frames, and the trajectory mapping of dynamic targets is used to analyze the movement direction and displacement characteristics of the light wave amplitude. The formula is used:

[0136]

[0137] Calculate the characteristic value of the light wave motion amplitude to obtain the light wave motion amplitude characteristic;

[0138] Among them, M represents the characteristic value of the amplitude of light wave motion, (x a ,y a ) and (x b ,y b ) represent the starting coordinates and ending coordinates of the target area in the continuous frames, θ1 and θ2 are the angles of the light wave motion direction in the continuous frames, and γ is the weight parameter;

[0139] The benefit of the formula is that, by combining the weighted calculation of the displacement distance and direction angle between consecutive frames, it can comprehensively reflect the motion characteristics of the light wave amplitude in space, taking into account both the displacement amplitude and the influence of the direction change through the weight parameter adjustment, thereby improving the accuracy of the analysis results;

[0140] M represents the comprehensive characteristics of the amplitude of light wave motion, which is obtained by weighted correction calculation of displacement distance and direction change. a ,y a ) and (x b ,y b) represent the initial and end coordinates of the target area in the continuous frames, which are obtained by marking the center point of the light wave amplitude distribution. θ1 and θ2 are the direction angles of the light wave amplitude in the continuous frames, which are obtained by fitting the trajectory points of the continuous frames. γ is a weight parameter used to balance the influence of the direction angle change, which is dynamically adjusted according to the importance of the changes in different directions.

[0141] Parameter value: x a =2.5,y a =3.0,x b =4.0,y b =5.5, θ1=30°, θ2=45°

[0142] , γ=0.8;

[0143] Calculate M:

[0144]

[0145] The result shows that the comprehensive characteristic value of the light wave motion amplitude is 14.92, which reflects the spatial displacement amplitude and direction change characteristics of the target area between consecutive frames, and can be used for subsequent trajectory analysis and motion description.

[0146] Call the light wave motion trajectory feature, combine the spatial displacement of the light wave with the motion direction, and match it with the trajectory parameters of the dynamic target. By analyzing the difference in the trajectory in the matching area, the dynamic motion characteristics of the target area are extracted, and a motion vector set of the target area is generated;

[0147] When the spatial displacement and motion direction of the light wave are matched with the trajectory parameters of the dynamic target, it is necessary to first standardize the trajectory characteristics of the light wave, extract the position parameters through the spatial distribution of the trajectory points, use the coordinate offset to represent the dynamic change of each trajectory point, perform the same discretization processing on the dynamic target trajectory parameters, perform point fitting on the spatial position parameters of the trajectory points, generate a discrete trajectory point set, match the positions of the light wave trajectory points with the dynamic target trajectory points, map the positions of the trajectory points of the two, calculate the spatial distance and direction offset difference between the trajectory points, and determine the overlapped part of the light wave motion trajectory and the dynamic target trajectory. Further, through the trajectory point matching results, the overall characteristics of the two trajectories are correlated and analyzed, including the distribution of the motion direction of the trajectory and the change of the trajectory curvature, etc., extract the motion vector characteristics of the target area, and finally integrate the vector set of the trajectory to generate the motion vector set of the target area.

[0148] See also Figure 8 ,The specific steps for obtaining the dynamic behavior recognition results of the monitoring target trajectory are:

[0149] Based on the motion vector set of the target area, the vector points are arranged in chronological order, the spatial displacement and time interval of adjacent vector points are identified, the displacement and time change ratios are compared, the dynamically changing vector points are screened, and a trajectory change node set is generated;

[0150] Each motion vector point is sorted according to the time axis, and the spatial displacement and the corresponding time interval between adjacent vector points are extracted in turn. The straight-line distance between adjacent vector points is calculated as the spatial displacement, and the time difference between the vector points is extracted in combination with the time axis data to calculate the displacement and time change ratio. For vector points with abnormal displacement change ratio or obvious jumps, their change trends are further analyzed to screen out vector points with significant dynamic change characteristics. In the screening process, points with low continuity or irregular changes are eliminated according to the cumulative deviation of spatial displacement and time ratio, and dynamic change points reflecting the characteristics of motion trajectory are retained. The data of the integrated points are used to generate a trajectory change node set, which provides a basis for trajectory dynamic analysis in holographic image monitoring.

[0151] Based on the trajectory change node set, the spatial distance and direction angle of the displacement path between adjacent nodes are identified, the path is converted into a curve for trajectory fitting, the continuity and local change rate of the fitting curve are parametrically analyzed, and the trajectory fitting result is generated;

[0152] The path length is calculated by extracting the node coordinates, and the direction angle characteristics are derived by combining the displacement vectors between the nodes. The data points of each path are converted into a continuous curve, and the motion trajectory curve is generated by serial fitting of the node coordinates. For the fitted trajectory curve, the continuity and local change rate of the curve are analyzed in turn, and the smoothness and dynamic change characteristics of the trajectory curve are quantified. In the parametric analysis, the local slope and second-order change rate of the curve are calculated, and the mutation position of the trajectory curve is marked to evaluate the integrity and change characteristics of the curve fitting. The above analysis results are integrated to generate the trajectory fitting result, which lays the foundation for the subsequent target dynamic behavior recognition.

[0153] Based on the trajectory fitting results, the speed and direction changes in the time series are analyzed in segments, the key positions of direction changes are marked, the action nodes are extracted, and the dynamic behavior recognition results of the monitored target trajectory are generated;

[0154] The trajectory is divided into several sections according to the time nodes, and the data characteristics of speed and direction changes are extracted section by section. The average speed and acceleration of each trajectory are calculated by fitting the data points of the curve, and the turning position is marked in combination with the change of the direction angle. In the process of turning marking, the time series of direction changes is analyzed for continuity, and the key nodes where significant directional deviations occur are screened. Combined with the comprehensive characteristics of speed and direction changes, the action nodes on the trajectory are extracted, and the dynamic change points reflecting the significant behavioral characteristics of the monitored target are marked. The action node data is integrated to generate the dynamic behavior recognition results of the monitored target trajectory, providing intelligent analysis capabilities for the holographic imaging monitoring system and supporting the accurate recognition and prediction of behavioral patterns.

[0155] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent video surveillance system based on holographic imaging, characterized in that: The system comprises: The multispectral imaging module is based on a multispectral holographic imaging device, which captures light wave amplitude, light wave frequency and light wave phase information, extracts infrared, ultraviolet and visible spectrum data, analyzes light wave energy and spatial distribution information, and generates a light wave energy distribution data set; The abnormal energy correction module analyzes the high value of light wave energy and the fluctuation amplitude of the partition based on the light wave energy distribution data set, matches the abnormal high value distribution area, identifies the light wave abnormal distribution area, obtains the abnormal energy identification data set, adjusts the distribution parameters in combination with the light wave characteristic differences in the monitoring image, and generates the corrected energy distribution result; The multi-spectral dynamic monitoring module extracts the light wave frequency characteristics of the frequency change area based on the corrected energy distribution result, analyzes the light wave frequency change area in combination with the dynamic target characteristics in the monitoring picture frame, generates the frequency change distribution result, matches the light wave characteristics and the video monitoring target characteristics in the area, locates the multi-spectral dynamic feature area, and obtains the multi-spectral dynamic area information; The behavior trajectory analysis module analyzes the spatial displacement and movement direction of the light wave amplitude in continuous frames based on the multi-spectral dynamic area information, generates a set of motion vectors of the target area in combination with the spatial change trajectory, performs time series analysis on the motion vector points, extracts trajectory change nodes, and generates dynamic behavior recognition results of the monitored target trajectory through node connection and trajectory fitting.

2. The intelligent video surveillance system based on holographic imaging according to claim 1 is characterized in that: The steps for acquiring the light wave energy distribution data set are specifically as follows: Based on multispectral holographic imaging equipment, the light wave amplitude, frequency and phase information are captured, infrared, ultraviolet and visible spectrum data are extracted, and the basic network structure of light wave energy distribution is established; The spatial distribution information of the basic network structure of light wave energy distribution is used, combined with the band weight parameters corresponding to the infrared, ultraviolet and visible spectrum data, to calculate the regional light wave energy value using the formula: Generate regional light wave energy analysis table; Among them, E represents the regional light wave energy value, A i represents the amplitude energy of the i-th light wave, P i represents the phase energy of the i-th light wave, W i represents the band weight parameter corresponding to the i-th light wave, x i Represents the horizontal axis of spatial distribution, y i represents the ordinate of spatial distribution, K is the adjustment coefficient, and n represents the total number of light waves; Based on the regional light wave energy analysis table, the light wave energy values ​​of the differentiated areas are compared, the energy distribution characteristics are analyzed, the spatial distribution trend of the light wave energy and the energy difference information are identified, and the light wave energy distribution data set is generated.

3. The intelligent video surveillance system based on holographic imaging according to claim 2 is characterized in that: The steps for obtaining the abnormal energy recognition data set are specifically as follows: Calling the light wave energy distribution data set, dividing the light wave energy value into intervals by analyzing the light wave energy value and the partition fluctuation amplitude in the area, marking the spatial distribution coordinates of the light wave energy value area, and generating a light wave energy high value spatial distribution set; Based on the high-value spatial distribution set of light wave energy, the energy difference in the high-value area is analyzed, and combined with the partition fluctuation amplitude parameter, the formula is adopted: Match the abnormal high energy value area to generate abnormal high value matching results; Among them, D represents an abnormally high matching value, E j represents the light wave energy value in the jth region, μ represents the mean value of the energy value, V j represents the fluctuation range of the jth region, σ represents the standard deviation of the fluctuation range, β is the smoothing coefficient, and m represents the number of regions; The abnormal high value matching result is called, and the abnormal matching value of the area is compared, and combined with the spatial distribution coordinates, the abnormal distribution area of ​​the light wave is identified, the light wave energy and spatial information in the abnormal area are extracted, and the abnormal energy identification data set is generated.

4. The intelligent video surveillance system based on holographic imaging according to claim 3 is characterized in that: The steps for obtaining the corrected energy distribution result are specifically as follows: Based on the abnormal energy identification data set, the original data of energy values ​​are extracted, grouped and processed according to the differentiated energy intervals in the data, the frequency of energy values ​​in each interval is counted, and the energy distribution characteristic parameters are obtained by quantifying the cumulative total amount of frequency data and combining the distribution deviation ratio of the energy values ​​in the interval; Based on the energy distribution characteristic parameters, light wave characteristic data of the monitoring image is collected, light wave intensity values ​​are allocated to corresponding distribution positions according to energy intervals, the proportional weights of intensity and interval parameters in the light wave data are adjusted, the frequency range of the energy characteristic parameters is allocated, and the light wave distribution parameters after adjustment are obtained; Based on the adjusted distribution parameters of the light waves, the energy values ​​are redistributed according to the interval frequencies, the weight ratios of the energy intervals are integrated, the energy value distribution is corrected according to the cumulative contribution of the light waves in the intervals, and a corrected energy distribution result is constructed.

5. The intelligent video surveillance system based on holographic imaging according to claim 4 is characterized in that: The steps for obtaining the frequency change distribution result are specifically as follows: Based on the corrected energy distribution result, the energy distribution is decomposed into a frequency change sequence according to the time series, the frequency change rate in the time series is analyzed, and the frequency change area of ​​the light wave is identified in combination with the statistical parameters of the frequency amplitude change range, and the frequency change area marking result is generated; The frequency change region marking result is called, the light wave frequency characteristics in the frequency change region are analyzed, and the frequency change characteristic value in the frequency change region is calculated using the formula: Generate frequency variation characteristic analysis results; Among them, F c Represents the frequency change characteristic value within the frequency change area, f k is the light wave frequency value at the kth time point, Δa k is the frequency amplitude change at the kth time point, Q k is the weight parameter, Δf k is the frequency change at the kth time point, α is the smoothing parameter; The frequency change characteristic analysis result is called, the frequency change region characteristic and the dynamic target feature are correlated and analyzed, the spatial position of the dynamic target and the intersection of the frequency change region are matched, and the frequency change distribution result is generated.

6. The intelligent video surveillance system based on holographic imaging according to claim 5 is characterized in that: The steps for obtaining the multi-spectral dynamic area information are specifically as follows: Based on the frequency change distribution results, the data are grouped according to the spatial region position, the amplitude and change trend of each group of frequency fluctuations are analyzed, the frequency dynamic distribution characteristics of the time axis data are identified, and the distribution correlation parameters between the frequency and the region are integrated to obtain the frequency and region matching data; Based on the frequency and area matching data, extract the intensity change characteristics of the differentiated areas in the light wave characteristic data, analyze the dynamic range of the light wave intensity on the time axis, compare the distribution of each set of intensity data with the spatial area, adjust the proportional relationship of the intensity distribution in the light wave characteristic, and generate a light wave and target area matching parameter set; Based on the light wave and target area matching parameter set, the spatial position of the multispectral feature is located in the monitoring data, the distribution data of the light wave characteristics in the time axis and the spatial area are analyzed, the multispectral feature frequency distribution weight of the feature area is screened, and the multispectral dynamic area information is obtained.

7. The intelligent video surveillance system based on holographic imaging according to claim 6 is characterized in that: The steps of obtaining the motion vector set of the target area are specifically as follows: Based on the multi-spectral dynamic area information, the time series of the light wave amplitude of the continuous frames is called, the spatial displacement of the light wave amplitude is analyzed, and the spatial change amount and directional change amount of the regional light wave amplitude between the continuous frames are analyzed to obtain the displacement distribution characteristics of the light wave amplitude; The displacement distribution characteristics of the light wave amplitude are called, combined with the spatial change trajectory of continuous frames, and the movement direction and displacement characteristics of the light wave amplitude are analyzed through trajectory mapping of dynamic targets, using the formula: Calculate the characteristic value of the light wave motion amplitude to obtain the light wave motion amplitude characteristic; Among them, M represents the characteristic value of the amplitude of light wave motion, (x a ,y a ) and (x b ,y b ) represent the starting coordinates and ending coordinates of the target area in the continuous frames, θ1 and θ2 are the angles of the light wave motion direction in the continuous frames, and γ is the weight parameter; The light wave motion trajectory feature is called, the spatial displacement of the light wave is combined with the motion direction, and matched with the trajectory parameters of the dynamic target. By analyzing the difference in the trajectory in the matching area, the dynamic motion feature of the target area is extracted, and a motion vector set of the target area is generated.

8. The intelligent video surveillance system based on holographic imaging according to claim 7 is characterized in that: The steps for obtaining the monitoring target trajectory dynamic behavior recognition result are specifically as follows: Based on the motion vector set of the target area, the vector points are arranged in time order, the spatial displacement and time interval of adjacent vector points are identified, the displacement and time change ratio are compared, the dynamically changing vector points are screened, and a trajectory change node set is generated; Based on the trajectory change node set, the spatial distance and direction angle of the displacement path between adjacent nodes are identified, the path is converted into a curve for trajectory fitting, the continuity and local change rate of the fitting curve are parameterized and analyzed, and a trajectory fitting result is generated; Based on the trajectory fitting results, the speed and direction changes in the time series are analyzed in segments, the key positions of the direction changes are marked, the action nodes are extracted, and the dynamic behavior recognition results of the monitoring target trajectory are generated.

Images